Annotation of complex mass spectra by multi-layered analysis

Highlights: • Unique Multi-layered approach to complex electrospray spectrum annotation. • LCMS data reduction via feature characterization for metabolomics and compound ID. • Identification of unexpected features, e.g. Ca, Al, Fe, Ba adducts. • Annotation of spectra comprising multiple compounds. • Software tool written in Python 3 with Jupyter Notebook framework. Annotating electrospray small molecule mass spectra remains a challenging problem due to the multiple processes occurring during ionization. Although an [M+H]⁺ is often present, ions can be formed by reactions with other cations, background compounds and co-eluting species, and by in-source fragmentation. Even single analytes can produce multiple ion forms, many of which remain unidentified and may appear to be different species, affecting reproducibility, quantification and precursor selection in DDA experiments. Annotation usually compares differences between peaks to known adducts and losses but fails if key peaks are missing or if the peaks are from unexpected adducts. Further, isotopes are often assumed to be due to ¹³C and removed prior to analysis which can leave 'orphan' peaks if unusual elements are present. Here we describe an alternative multi-layered approach (MLA) which successively matches spectra to calculated target ion lists and reprocesses the residual ions. This allows the analyst to focus on the unknown ions and to progressively increase target list complexity since explained ions are removed. Target ion lists can be calculated from expected or observed masses and potential adducts or can be pre-defined lists, for example common contaminants. Using this approach on spectra of known standards we identified adducts with Ca, Al, Fe, Ba and possibly Mg and Sr. We also detected several compounds and adducts in a spectrum of co-eluting species from an LC-MS analysis.